1. Field of the Invention
The present invention relates to image data processing techniques and in particular to a watermarking and watermark detecting system and method for a scaled-up/down image.
2. Description of the Prior Art
With wide-spreading digital satellite broadcasting, the Internet, and DVD (Digital Versatile Disk), etc. in recent years, digital images are becoming easily accessible to users. Since the quality of digital images does not deteriorate when they are copied, copyright protection has been an important issue.
For protection of video contents, motion-picture or still-picture contents and the like, there has been proposed a digital watermarking system. More specifically, at a contents provider, watermark data is embedded into DCT (discrete cosine transform) coefficient domain of compressed image data. At a user terminal, it is checked whether a watermark is embedded in the input image data. When the watermark has been detected from the input image data, the watermark detection result is displayed and thereby copying can be effectively prevented (see Japanese Patent Application Unexamined Publication No. 10-155151).
On the other hand, with the widespread use of DVD recorders, there has been a growing trend to permit a user to copy the content of a DVD only a limited number of times, usually only once. The limited number of times is determined by a copy control watermark previously embedded in the image content of the DVD. In the case of a copy-once watermark indicating that a copy is permitted only once, an appended-type watermark is embedded in the image content when the copy has been performed. When the appended-type watermark is detected from an input image content, the DVD recorder prohibits any copy. Such a copy protection system effectively prevents wide-spreading illegal copies.
Accordingly, reliable detection of such a watermark is a very important issue to enhance the copy protection.
In some cases, however, an input image having a copy control watermark embedded therein may be scaled up/down by using a video editor and the like. It is known that such scaling makes detection of an appended-type watermark difficult. For example, when a watermarked image has been scaled by a horizontal factor Sx and vertical factor Sy, it is necessary to scale an appended-type watermark by the same factor before embedding it into the scaled watermarked image. When the scaling factor of the scaled appended-type watermark is equal to that of the scaled watermarked image, the appended-type watermark can be detected with reliability even if the scaled watermarked image having the scaled appended-type watermark embedded therein is scaled again.
However, in the case where the scaled watermarked image having the scaled appended-type watermark embedded therein is scaled again, the original appended-type watermark is eventually scaled two times. This deteriorates the detection accuracy of the appended-type watermark. On the other hand, when the appended-type watermark is embedded into the scaled image, it is necessary to scale down the appended-type watermark two times. This increases the size of a necessary program.